- Fix a bug where CRI implementations that use cAdvisor stats provider (CRI-O) don't evict pods when their logs exceed ephemeral storage limit. ([#108115](https://github.com/kubernetes/kubernetes/pull/108115), [@haircommander](https://github.com/haircommander)) [SIG Node]

Kubernetes cgroup manager uses `memory.limit_in_bytes` in v1 cgroups to limit the memory capacity for a container, and uses `oom_scores` to recommend an order for killing container processes if an out-of-memory event occurs. This implementation has shortcomings: for `Guaranteed` pods, memory can not be fully reserved, and the page cache is at risk of being recycled. For `Burstable` pods, overcommitting memory (setting `request` less than `limit` ) could increase the risk of a container being killed when the...

labels.backup_configuration=Backup
labels.backup_name=Name
labels.backup_bulk_file=Bulk File
labels.backup_button_upload=Upload
labels.process_time_is_exceeded=The search processing time has exceeded the limit. The displayed results may be partial.
labels.user_given_name=Given Name
labels.givenName=Given Name
labels.user_surname=Surname
labels.surame=Surname
labels.user_mail=Mail
labels.mail=Mail

		Hidden: true,
		EnvVar: "MINIO_CROSSDOMAIN_XML",
	},
	cli.StringFlag{
		Name:   "memlimit",
		Usage:  "set global memory limit per server via GOMEMLIMIT",
		Hidden: true,
		EnvVar: "MINIO_MEMLIMIT",
	},
	cli.IntFlag{
		Name:   "send-buf-size",
		Value:  4 * humanize.MiByte,
		EnvVar: "MINIO_SEND_BUF_SIZE",
		Hidden: true,
	},

//
// Furthermore if a has more than dataScannerCompactAtChildren recursive children (uncompacted folders)
// the tree will be recursively scanned and the branches with the least number of objects will be
// compacted until the limit is reached.
//
// This ensures that any branch will never contain an unreasonable amount of other branches,
// and also that small branches with few objects don't take up unreasonable amounts of space.

	if err != nil {
		return result, toObjectErr(err, bucket, object, uploadID)
	}

	uploadIDPath := er.getUploadIDDir(bucket, object, uploadID)
	if partNumberMarker < 0 {
		partNumberMarker = 0
	}

	// Limit output to maxPartsList.
	if maxParts > maxPartsList {
		maxParts = maxPartsList
	}

	// Populate the result stub.
	result.Bucket = bucket
	result.Object = object
	result.UploadID = uploadID

- Limit number of instances in a single update to GCE target pool to 1000. ([#87881](https://github.com/kubernetes/kubernetes/pull/87881), [@wojtek-t](https://github.com/wojtek-t)) [SIG Cloud Provider, Network and Scalability]

* Send recycle events from pod to pv. ([#27714](https://github.com/kubernetes/kubernetes/pull/27714), [@jsafrane](https://github.com/jsafrane))
* Limit the number of names per image reported in the node status ([#32914](https://github.com/kubernetes/kubernetes/pull/32914), [@yujuhong](https://github.com/yujuhong))

            final byte[] buf = buffer.array();
            int len;
            int amount = 0;
            while ((len = in.read(buf)) != -1) {
                buffer.limit(len);
                channel.write(buffer, amount);
                buffer.clear();
                amount += len;
            }
            out.flush();
            return amount;

	mac := hmac.New(sha256.New, d.objectEncryptionKey) // derive part encryption key from part ID and object key
	mac.Write(partIDbin[:])
	partEncryptionKey := mac.Sum(nil)

	// Limit the reader, so the decryptor doesn't receive bytes
	// from the next part (different DARE stream)
	encLenToRead := d.parts[d.partIndex].Size - d.partEncRelOffset